Ignaz Semmelweis had blood on his hands

Actually. Not a metaphor.

This is a brilliant translation of a paper Semmelweis wrote more than a hundred years ago. It teaches us an enormous amount about science, about choosing to see, and about the essential work of figuring things out.

This essay will help you understand what a theory actually is, what happens when the world presents us with a beautifully delineated problem, and what we might choose to do about it.

In today’s tl;dr world, you might be tempted to read a quick summary and then move onto a funny cat video. I hope you’ll slow down for ten minutes and try to imagine what it was like to be Semmelweis, to be surrounded by mysterious disease and death, to spend a moment or two looking into the numbers… and then, if you’re still along for the ride, to imagine what you would have done once you realized that it was you, your hands, your actions, that killed so many mothers and their children…

It’s a mystery, a puzzle, a heroic journey of science.

We need more of this. And each of us is capable of it. Figure it out. Do the math. Spin a theory. Prove it. Repeat.

Summary: Two hospitals, identical in almost every way, except in one of them, mothers and their babies are dying. This goes on for years. Sometimes it’s people in beds next to each other, other times not. What causes the disease? Why in one hospital and not another? Why doesn’t it strike mothers who give birth just before arriving at the hospital…

The punchline becomes clear about halfway through. Try to imagine what it took for him to figure this out when Louis Pasteur hadn’t yet done his work, and medicine was basically at the level of superstition…

Source: Excerpted from Ignaz Semmelweis, The Etiology, Concept, and Prophylaxis of Childbed. Fever, trans. K. Codell Car­ter. Madison, The University of Wisconsin Press, 1983.


Ignaz Semmelweis


Medicine’s highest duty is saving threatened human life, and obstetrics is the branch of med­icine in which this duty is most obviously ful­filled. Frequently it is necessary to deliver a child in transverse lie. Mother and child will probably die if the birth is left to nature, while the obstetrician’s timely helping hand, almost painlessly and taking only a few minutes, can save both.

1 was already familiar with this prerogative of obstetrics from the theoretical lectures on the specialty. I found it perfectly confirmed as I had the opportunity to learn the practical aspects of obstetrics in the large Viennese maternity hos­pital. But unfortunately the number of cases in which the obstetrician achieves such blessings vanishes in comparison with the number of vic­tims to whom his help is of no avail. This dark side of obstetrics is childbed fever. Each year I saw ten or fifteen crises in which the salvation of mother and child could be achieved. I also saw many hundreds of maternity patients treated unsuccessfully for childbed fever. Not only was therapy unsuccessful, the etiology seemed defi­cient. The accepted etiology of childbed fever, on the basis of which I saw so many hundreds of maternity patients treated unsuccessfully, can­not contain the actual causal factor of the dis­ease.

The large gratis Viennese maternity hospital is divided into two clinics; one is called the first, the other the second. By Imperial Decree of 10 October 1840, Court Commission for Educa­tion Decree of 17 October 1840, and Adminis­trative Ordinance of 27 October 1840, all male students were assigned to the first clinic and all female students to the second. Before this time the admission of maternity patients was reg­ulated as follows: Monday afternoon at four o’clock admissions began in the first clinic and continued until Tuesday afternoon at four. Ad­missions then began in the second clinic and continued until Wednesday afternoon at four o’clock. At that time admissions were resumed in the first clinic until Thursday afternoon, etc. On Friday afternoon at four o’clock admissions began in the first clinic and continued through forty-eight hours until Sunday afternoon, at which time admissions began again in the sec­ond clinic. Admissions alternated between the two clinics through twenty-four hour periods, and only once a week did admissions continue in the first clinic for forty-eight hours. Thus the first clinic admitted patients four days a week, whereas the second clinic admitted for only three days. The first clinic, thereby, had fifty-two more days of admissions [each year] than the second.

From the time the first clinic began training only obstetricians until June 1847, the mortality rate in the first clinic was consistently greater than in the second clinic, where only midwives were trained. Indeed, in the year 1846, the mor­tality rate in the first clinic was five times as great as in the second, and through a six-year period it was, on the average, three times as great. This is shown in Table I.

The difference in mortality between the clinics was actually larger than the table sug­gests, because occasionally, for reasons to be examined later,’ during times of high mortality all ill maternity patients in the first clinic were transferred to the general hospital. When these patients died, they were included in the mor­tality figures for the general hospital rather than for the maternity hospital. When the transfers were undertaken, the reports show re­duced mortality, since only those who could not be transferred because of the rapid course of their illness were included. In reality, many ad­ditional victims should be included. In the sec­ond clinic such transfers were never under­taken. Only isolated patients were transferred whose condition might endanger the other pa­tients. . . .

. . . What is the origin, then, of the difference in mortality between the clinics? Hyperinosis [excessive fibrin in the blood], hydremia [excessive water in the blood], plethora [an ex­cessive quantity of blood], disturbances caused by the pregnant uterus, stagnation of the cir­culation, inopexia [spontaneous coagulation of the blood], delivery itself, decreased weight caused by the emptying of the uterus, pro­tracted labor, wounding of the inner surface of the uterus in delivery, imperfect contractions, faulty involutions of the uterus during mater­nity, scanty and discontinued secretion and ex­cretion of lochia [a vaginal discharge during the first few weeks after delivery], the weight of secreted milk, death of the fetus, and the indi­viduality of patients are causes to which may be ascribed much or little influence in the genera­tion of childbed fever. But in both clinics these must be equally harmful or harmless and they cannot, therefore, explain the appalling differ­ence in mortality between the clinics.

While I was still unable to find a cause for the increased mortality rate in the first clinic, I be­came aware of other inexplicable circumstances. Those whose period of dilation was extended over twenty-four hours or more almost invari­ably became ill either immediately during birth or within the first twenty-four or thirty-six hours after delivery. They died quickly of rapidly developing childbed fever. An equally extended period of dilation in the second clinic did not prove dangerous. Because dilation was usually extended during first deliveries, those delivering for the first time usually died. I often pointed out to my students that because these blossoming, vigorously healthy young women had extended periods of dilation, they would die quickly from puerperal fever either during delivery or immediately thereafter. My prog­noses were fulfilled. I did not know why, but I often saw it happen. This circumstance was in­explicable, since it was not repeated in the sec­ond clinic. I speak here of the period of dila­tion, not of delivery; thus the trauma of delivery is not under consideration.

Not only these mothers but also their new­born infants, both male and female, died of childbed fever. I am not alone in speaking of puerperal fever of the newborn.2 With the ex­ception of the genital areas, the anatomical le­sions in the corpses of such newborn infants are the same as the lesions in the corpses of women who die of puerperal fever. To recognize these findings as the consequence of puerperal fever in maternity patients but to deny that identical findings in the corpses of the newborn are the consequence of the same disease is to reject pathological anatomy.

But if the maternity patients and the new­born die from the same disease, then the etiology that accounts for the deaths of the mothers must also account for the deaths of the newborn. Since the difference in mortality be­tween the maternity patients in the two clinics was reflected in the mortality rates for the new­born, the accepted etiology for childbed fever no more accounts for the deaths of the newborn than for the deaths of the maternity patients. Table 2 gives the mortality rates of the newborn at the two clinics.

Because their mothers died or were otherwise unable to nurse, many of the newborn were sent directly to the foundling home. Later we will consider their fate.

The occurrence of childbed fever among the newborn can be explained in two ways. Child-bed fever may be caused by factors operating on the mother during the intrauterine life of the fetus, and the mother can then impart the dis­ease to the infant. Alternatively, the causes may affect the infant itself after birth, in which case the mother may or may not be affected. Thus the infant dies, not because the disease has been imparted, as in the first case, but rather because childbed fever originates in the infant itself. If the mother imparts childbed fever to the infant during intrauterine life, then the difference in infant mortality between the two clinics cannot be explained by the accepted etiology, because this etiology inadequately explains the origin of the disease in mothers. If the cause of childbed fever operates directly on the infant indepen­dently of the mother, then it is still impossible for the accepted etiology to explain the differ­ence in infant mortality rates. [Given the ac­cepted theories], one would expect the mor­tality in the second clinic to have been either equal to or greater than that of the first. Of course, many of the causal factors that purport­edly explain childbed fever among maternity patients are simply impossible with regard to infants — infants would not, in all probability, fear the evil reputation of the first clinic, their modesty would not be offended by having been delivered in the presence of men, etc.

Childbed fever is defined as a disease charac­teristic of and limited to maternity patients, for whose origin the puerperal state and a specific causal moment are necessary.4 Thus when this cause operates on a person who is predisposed by the puerperal state, childbed fever results. However, if this same cause operates on persons who are not puerperae, some disease other than puerperal fever is generated. For example, some believe that maternity patients in the first clinic, knowing of the countless deaths occur­ring there each year, are so frightened that they contract the disease. Thus the disposing factor is the puerperal state, and the precipitating fac­tor is fear of death. We can assume that many soldiers engaged in murderous battle must also fear death. However, these soldiers do not con­tract childbed fever, because they are not puer­perae and so they lack the disposing factor.

If an individual is openly examined for the instruction of males, her modesty is offended and, because she is predisposed by the puer­peral state, she contracts childbed fever. But female modesty can be offended in many ways, and if the offended young woman is not in the puerperal state, she does not contract childbed fever because she is not predisposed. Something else occurs; for example, she may swoon. Chill­ing brings childbed fever in puerperae, but in other persons it causes rheumatic fever. In puerperae, mistakes in diet induce childbed fever. In others, similar mistakes cause only gas­tric fever.

Becoming convinced that childbed fever is not restricted to puerperae and that it can begin during birth or even in pregnancy, one may ignore the puerperal state and focus on the unique composition of the blood during preg­nancy. But even if we adopt such an approach, what predisposes the newborn to puerperal dis­ease? Surely not the puerperal condition of their genitals. Do both male and female have the blood composition uniquely characteristic of pregnancy? The occurrence of childbed fever among the newborn shows that the very conception of puerperal fever is erroneous.

Because Vienna is so large, women in labor often deliver on the street, on the glacis or in front of the gates of houses before they can reach the hospital. It is then necessary for the woman, carrying her infant in her skirts, and often in very bad weather, to walk to the mater­nity hospital. Such births are referred to as street births. Admission to the maternity clinic and to the foundling home is gratis, on the condition that those admitted be available for open instructional purposes, and that those fit to do so serve as wet nurses for the foundling home. Infants not born in the maternity clinic are not admitted gratis to the foundling home because their mothers have not been available for instruction. However, in order that those who had the intention of delivering in the ma­ternity hospital but who delivered on the way would not innocently lose their privilege, street births were counted as hospital deliveries. This, however, led to the following abuse: women in somewhat better circumstances, seeking to avoid the unpleasantness of open examination without losing the benefit of having their in­fants accepted gratis to the foundling home, would be delivered by midwives in the city and then be taken quickly by coach to the clinic where they claimed that the birth had occurred unexpectedly while they were on their way to the clinic. If the child had not been christened

As I have noted, women who delivered on the street contracted childbed fever at a signifi­cantly lower rate than those who delivered in the hospital. This was in spite of the less favor­able conditions in which such births took place. Of course, in most of these cases delivery oc­curred in a bed with the assistance of a midwife. Moreover, after three hours our patients were obliged to walk to their beds by way of the glass-enclosed passageway. However, such inconve­nience is certainly less dangerous than being delivered by a midwife, then immediately hav­ing to arise, walk down many flights of stairs to the waiting carriage, travel in all weather condi­tions and over horribly rough pavement to the maternity hospital, and there having to climb up another flight of stairs. For those who really gave birth on the street, the conditions would have been even more difficult.

To me, it appeared logical that patients who experienced street births would become ill at least as frequently as those who delivered in the clinic. I have already expressed my firm convic­tion that the deaths in the first clinic were not caused by epidemic influences but by endemic and as yet unknown factors, that is, factors whose harmful influences were limited to the first clinic. What protected those who delivered outside the clinic from these destructive un­known endemic influences? In the second clinic, the health of the patients who underwent street births was as good as in the first clinic, but there the difference was not so striking, since the health of the patients was generally much better.

This would be the place to exhibit a table showing that the mortality rate among those who delivered on the street was lower than among those who delivered in the first clinic. While I had access to the records of the first clinic I felt that such a table was unnecessary because no one denied these facts. Thus I ne­glected to complete a table. Later when I was no longer assistant, these facts were denied, as was the existence of a significant difference in mor­tality between the clinics. Because of Table I, however, this difference is undeniable. In 1848 Professor [Josef] Skoda proposed that the fac­ulty of the Viennese medical school nominate a commission that, among other things, would construct such a table. The proposal was adopted by a great majority, and the commis­sion was immediately named. However, as a re­sult of protests by the Professor of Obstetrics, higher authorities intervened and the commis­sion was unable to begin its activity.

In addition to those who delivered on the street, those who delivered prematurely also be­came ill much less frequently than ordinary pa­tients. Those who delivered prematurely were not only exposed to all the same endemic influ­ences as patients who went full-term, they also suffered the additional harm of whatever caused the premature delivery. Under these cir­cumstances, how could their superior health be explained? One explanation was that the earlier the birth, the less developed the puerperal con­dition and therefore the smaller the predisposi­tion for the disease. Yet puerperal fever can begin during birth or even during pregnancy; indeed, even at these times it can be fatal. The better health of patients who delivered pre­maturely in the second clinic conformed to the general superior health of full-term patients in the clinic.

Patients often became ill sporadically. One diseased patient would be surrounded by healthy patients. But very often whole rows would become ill without a single patient in the row remaining healthy. The beds in the mater­nity wards were arranged along the length of the rooms and were separated by equal spaces.

Depending on their location, rooms in the clinic extended either north-south or east-west. If pa­tients in beds along the north walls became ill we were often inclined to regard chilling as a significant factor. However, on the next occa­sion those along the south wall would become ill. Many times those on the east and west walls would become diseased. Often the disease spread from one side to the other, so that no one location seemed better or worse. How could these events be explained, given that the same patterns did not appear in the second clinic where one encountered the disease only spo­radically?

It was my firm conviction that childbed fever was not contagious and did not spread from bed to bed. Later we will consider the proof for this conviction. For now, it is sufficient to note that the disease appeared only sporadically in the second clinic. If childbed fever were contagious, from the sporadic cases whole rows would be­come ill as the disease spread from bed to bed.

The authorities did not remain indifferent to the disturbing difference in mortality rates be­tween the two clinics. Commissions repeatedly investigated and conducted hearings to deter­mine the cause of the difference, and to decide whether it was possible to save a larger number of those patients who became ill. To achieve this last goal, from time to time all the diseased patients were transferred to the general hospi­tal. But in spite of the change in physicians, rooms, and medical procedures, etc., the pa­tients died almost without exception. The com­missions would conclude that the cause of the great mortality rate was one or another or several of the endemic factors previously discussed. Various suitable measures were adopted, but none succeeded in bringing the death rate within the limits set by the second clinic. The failure of these measures proved that the factors identified were not, in fact, the rele­vant causes.

Toward the end of 1846 an opinion prevailed in one commission that the disease originated from damage to the birth canal inflicted during the examinations that were part of the instruc­tional process. However, since similar examina­tions were part of the instruction of midwives, the increased incidence of disease in the clinic for physicians was made intelligible by assuming that male students, particularly foreigners, were too rough in their examinations. As a result of this opinion the number of students was re­duced from forty-two to twenty. Foreigners were almost entirely excluded, and examinations were reduced to a minimum. The mortality rate did decline significantly in December 1846, and in January, February, and March of 1847. But in spite of these measures, fifty-seven patients died in April and thirty-six more in May. This demonstrated to everyone that the view was groundless. To further the reader’s under­standing, Table 3 shows the mortality figures for 1846 and for the first five months of 1847. We will come back later to the fact that from December 1846 through March 1847 the mor­tality rate declined, and that it climbed back up again in April and May 1847.

Recommendations based on studies of the cause of the great mortality in the first clinic all involved one inexplicable contradiction: given the concept of an epidemic, and given that the commissions did not have the power to change the atmospheric-cosmic-terrestrial conditions of Vienna, they should have concluded that no remedies were possible. But they did not draw this conclusion, even though they considered the deaths an epidemic. What does one do to shorten the duration or to prevent the recur­rence of a cholera epidemic? They attributed the disease to one or more of the previously identified endemic causes. They did not, how­ever, identify it as an endemic disease, which would have been appropriate, but rather as an epidemic. In general, the unfortunate confu­sion in the concepts of epidemic and endemic disease delayed discovery of the true cause of childbed fever.

In classifying puerperal disease as epidemic or endemic, one must disregard entirely the number of patients who become ill or die. The cause of the illness or death determines whether the disease is epidemic or endemic. Epidemic puerperal fever is induced by atmo­spheric-cosmic-terrestrial influences; the con­cept of an epidemic does not stipulate whether one or one hundred persons become ill. If puerperal fever is caused by endemic factors — that is, by factors whose operation is limited to a specific location — then puerperal fever is en­demic, and it is immaterial whether one or one hundred individuals become ill. This follows from the concepts of epidemic and endemic disease. In classifying the disease one way or the other, however, the commissions did not con­sider the purported cause but only the number of cases. Because many patients became ill and died, it was identified as an epidemic.

I was convinced that the greater mortality rate at the first clinic was due to an endemic but as yet unknown cause. That the newborn, whether female or male, also contracted child-bed fever convinced me that the disease was misconceived. I was aware of many facts for which I had no explanation. Delivery with pro­longed dilation almost inevitably led to death. Patients who delivered prematurely or on the street almost never became ill, and this contra­dicted my conviction that the deaths were due to endemic causes. The disease appeared se­quentially among patients in the first clinic. Pa­tients in the second clinic were healthier, al­though individuals working there were no more skillful or conscientious in their duties. The disrespect displayed by the employees toward the personnel of the first clinic made me so miserable that life seemed worthless. Every­thing was in question; everything seemed inex­plicable; everything was doubtful. Only the large number of deaths was an unquestionable reality.

The reader can appreciate my perplexity dur­ing my first period of service when I, like a drowning person grasping at a straw, discon­tinued supine deliveries, which had been cus­tomary in the first clinic, in favor of deliveries from a lateral position. I did this for no other reason than that the latter were customary in the second clinic. I did not believe that the supine position was so detrimental that addi­tional deaths could be attributed to its use. But in the second clinic deliveries were performed from a lateral position and the patients were healthier. Consequently, we also delivered from the lateral position, so that everything would be exactly as in the second clinic.

I spent the winter of 1846–47 studying Eng­lish. I did this because my predecessor, Dr. Breit, resumed the position of assistant, and I wanted to spend time in the large Dublin mater­nity hospital. Then, at the end of February 1847, Dr. Breit was named Professor of Obstetrics at the medical school in Tubingen. I changed my travel plans and, in the company of two friends, departed for Venice on 2 March 1847. I hoped that Venetian art treasures would revive my mind and spirits, which had been so seriously affected by my experiences in the ma­ternity hospital.

On 20 March of the same year, a few hours after returning to Vienna, I resumed, with re­juvenated vigor, the position of assistant in the first clinic. I was immediately overwhelmed by the sad news that Professor [Jakob] Kolletschka, whom I greatly admired, had died in the in­terim.

The case history went as follows: Kolletschka, Professor of Forensic Medicine, often con­ducted autopsies for legal purposes in the com­pany of students. During one such exercise, his finger was pricked by a student with the same knife that was being used in the autopsy. I do not recall which finger was cut. Professor Kolletschka contracted lymphangitis and phle­bitis [inflammation of the lymphatic vessels and of the veins respectively] in the upper extremity. Then, while 1 was still in Venice, he died of bilateral pleurisy, pericarditis, peritonitis, and meningitis [inflammation of the membranes of the lungs and thoracic cavity, of the fibroserous sac surrounding the heart, of the membranes of the abdomen and pelvic cavity, and of the mem­branes surrounding the brain, respectively]. A few days before he died, a metastasis also formed in one eye. I was still animated by the art treasures of Venice, but the news of Kolletschka’s death agitated me still more. In this excited condition I could see clearly that the disease from which Kolletschka died was identical to that from which so many hundred maternity patients had also died. The maternity patients also had lymphangitis, peritonitis, peri­carditis, pleurisy, and meningitis, and metasta­ses also formed in many of them. Day and night I was haunted by the image of Kolletschka’s disease and was forced to recognize, ever more decisively, that the disease from which Kolletschka died was identical to that from which so many maternity patients died.

Earlier, I pointed out that autopsies of the newborn disclosed results identical to those ob­tained in autopsies of patients dying from child-bed fever. I concluded that the newborn died of childbed fever, or in other words, that they died from the same disease as the maternity patients. Since the identical results were found in Kolletschka’s autopsy, the inference that Kolletschka died from the same disease was confirmed. The exciting cause of Professor Kolletschka’s death was known; it was the wound by the autopsy knife that had been con­taminated by cadaverous particles. Not the wound, but contamination of the wound by the cadaverous particles caused his death. Kolletschka was not the first to have died in this way. I was forced to admit that if his disease was identical with the disease that killed so many maternity patients, then it must have originated from the same cause that brought it on in Kolletschka. In Kolletschka, the specific causal factor was the cadaverous particles that were introduced into his vascular system. I was com­pelled to ask whether cadaverous particles had been introduced into the vascular systems of those patients whom I had seen die of this iden­tical disease. I was forced to answer affirma­tively.

Because of the anatomical orientation of the Viennese medical school, professors, assistants, and students have frequent opportunity to con­tact cadavers. Ordinary washing with soap is not sufficient to remove all adhering cadaverous particles. This is proven by the cadaverous smell that the hands retain for a longer or shorter time. In the examination of pregnant or deliver­ing maternity patients, the hands, contaminated with cadaverous particles, are brought into contact with the genitals of these individuals, creat­ing the possibility of resorption. With resorp­tion, the cadaverous particles are introduced into the vascular system of the patient. In this way, maternity patients contract the same dis­ease that was found in Kolletschka.

Suppose cadaverous particles adhering to hands cause the same disease among maternity patients that cadaverous particles adhering to the knife caused in Kolletschka. Then if those particles are destroyed chemically, so that in ex­aminations patients are touched by fingers but not by cadaverous particles, the disease must be reduced. This seemed all the more likely, since I knew that when decomposing organic material is brought into contact with living organisms it may bring on decomposition.

To destroy cadaverous matter adhering to hands I used chlorina liquida. This practice be­gan in the middle of May 1847; I no longer remember the specific day. Both the students and I were required to wash before examina­tions. After a time I ceased to use chlorina liquida because of its high price, and I adopted the less expensive chlorinated lime. In May 1847, dur­ing the second half of which chlorine washings were first introduced, 36 patients died-this was 12.24 percent of 294 deliveries. In the re­maining seven months of 1847, the mortality rate was below that of the patients in the second clinic (see Table 4).

In these seven months, of the 1841 maternity patients cared for, 56 died (3.04 percent). In 1846, before washing with chlorine was intro­duced, of 4010 patients cared for in the first clinic, 459 died (11.4 percent). In the second clinic in 1846, of 3754 patients, 105 died (2.7 percent). In 1847, when in approximately the middle of May I instituted washing with chlo­rine, in the first clinic of 3490 patients, 176 died (5 percent). In the second clinic of 3306 pa­tients, 32 died (0.9 percent). In 1848, chlorine washings were employed throughout the year and of 3556 patients, 45 died (1.27 percent). In the second clinic in the year 1848, of 3219 pa­tients 43 died (1.33 percent). The mortality rates for the individual months of 1848 are shown in Table 5.

In March and August 1848 not a single pa­tient died. In January 1849, of 403 births 9 died (2.23 percent). In February, of 389 births 12 died (3.08 percent). March had 406 births, and there were 20 deaths (4.9 percent). On 20 March Dr. Carl Braun9 succeeded me as as­sistant.

As mentioned, the commissions identified various endemic factors as causes of the greater mortality rate in the first clinic. Accordingly, various measures were instituted, but none brought the mortality rate within that of the second clinic. Thus one could infer that the factors identified by the commissions were not causally responsible for the greater mortality in the first clinic. I assumed that the cause of the greater mortality rate was cadaverous particles adhering to the hands of examining obstetri­cians. I removed this cause by chlorine wash­ings. Consequently, mortality in the first clinic fell below that of the second. I therefore con­cluded that cadaverous matter adhering to the hands of the physicians was, in reality, the cause of the increased mortality rate in the first clinic. Since the chlorine washings were instituted with such dramatic success, not even the smallest additional changes in the procedures of the first clinic were adopted to which the decline in mor­tality could be even partially attributed. The instruction system for midwives is so instituted that pupils and instructors have less frequent occasion to contaminate their hands with ca­daverous matter than is the case in the first clinic. Thus, the unknown endemic cause of the horrible devastations in the first clinic was the cadaverous particles adhering to the hands of the examiners.

In order to destroy the cadaverous material, it was necessary that every examiner wash in chlo­rinated lime upon entry into the labor room. Because students in the labor room had no op­portunity to contaminate their hands anew, I believed one washing was sufficient. Because of the large number who gave birth each year in the first clinic, patients were seldom alone in the labor room; as a rule several were there simul­taneously. For purposes of instruction, those in labor were arranged and examined sequen­tially. I regarded it as sufficient that after each examination the hands were washed with soap and water only. Within the labor room, it seemed unnecessary for the hands to be washed with chlorine water between examinations. Once the hands had been cleaned of ca­daverous particles, they could not become con­taminated again.

In October 1847, a patient was admitted with discharging medullary carcinoma [cancer of the innermost parti of the uterus. She was assigned the bed at which the rounds were always initi­ated. After examining this patient, those con­ducting the examination washed their hands with soap only. The consequence was that of twelve patients then delivering, eleven died. The ichor from the discharging medullary car­cinoma was not destroyed by soap and water. In the examinations, ichor was transferred to the remaining patients, and so childbed fever multi­plied. Thus, childbed fever is caused not only by cadaverous particles adhering to hands but also by ichor from living organisms. It is necessary to clean the hands with chlorine water, not only when one has been handling cadavers but also after examinations in which the hands could become contaminated with ichor. This rule, originating from this tragic experience, was fol­lowed thereafter. Childbed fever was no longer spread by ichor carried on the hands of exam­iners from one patient to another.

A new tragic experience persuaded me that air could also carry decaying organic matter. In November of the same year, an individual was admitted with a discharging carious left knee. In the genital region this person was completely healthy. Thus the examiners’ hands presented no danger to the other patients. But the ich­orous exhalations of the carious knee com­pletely saturated the air of her ward. In this way the other patients were exposed and nearly all the patients in that room died. The reports of the first clinic indicate that eleven patients died in November and eight more in December. These deaths were largely due to ichorous ex­halations from this individual. The ichorous particles that saturated the air of the maternity ward penetrated the uteruses already lacerated in the birth process. The particles were re-sorbed, and childbed fever resulted. Thereafter, such individuals were isolated to prevent similar tragedies.

The maternity hospital in Vienna was opened on 16 August 1784. In the eighteenth century and in the early decades of the nineteenth cen­tury, medicine was concerned with theoretical speculation, and the anatomical foundations were neglected. Thus in 1822, of 3066 patients only 26 died (.84 percent). In 1841, after the Viennese medical school adopted an anatomical orientation, of 3036 patients 237 died (7.7 per­cent). In 1843 of 3060 patients 274 died (8.9 percent). In 1827, of 3294 patients 55 died (1.66 percent). In 1842 of 3287 patients 518 died (15.8 percent))° From 1784 until 1823, over a period of twenty-five years, less than 1 percent of the patients cared for in the mater­nity hospital died. This is shown in Table 6.

10 The figures for 1841, 1842, and 1843 are for the first clinic only, see Table I.

This table provides unchallengeable proof for my opinion that childbed fever originates with the spread of animal-organic matter. At the time when the educational system limited op­portunities for spreading decaying animal-organic matter, the patients cared for in the maternity hospital were much healthier.

As the Viennese medical school adopted an anatomical orientation, the health of the mater­nity patients worsened. When the number of births and of students became so great that one professor could not supervise the births and give instruction, the maternity hospital was di­vided into two clinics. At that time the same number of male and female students were as­signed to each clinic. On 10 October 1840, by imperial decree, all males were assigned to the first clinic and all female students to the second. I am not able to say in which year the maternity hospital was divided. Colleagues who taught obstetrics in the second clinic when male stu­dents were still admitted report that there was, at that time, no significant difference in mor­tality between the clinics. The consistently un­favorable health of patients in the first clinic dates from 1840, when all male students were assigned to the first clinic and all female stu­dents to the second. After what has been re­ported, it would be superfluous to explain these facts further.

Table 1 indicates the difference in mortality rates between the patients of the two clinics after the first was devoted exclusively to train­ing obstetricians and the second to training midwives. This would be the place to provide a similar table for the years during which female and male students were divided equally be­tween both clinics. It would show that during this time the mortality rate was not consistently larger in the first clinic. However, I do not have access to the necessary data. The reports were prepared in triplicate in both clinics. One copy remained in the institution; one copy was sent to the governmental administration. Those who now have these reports would do a service to science if they would release them to the pub­lic.” I possess the reports of both clinics only for 1840, when the male and female students were separated, and for the preceding year (see Table 7). The variation in mortality for both clinics can be traced to the activities of those in the process of becoming physicians. I was obstructed in disclosing this information be­cause at the time it was construed as a basis for personal denunciation.

Professor Skoda assigned various responsibil­ities to the above mentioned commission of the Viennese medical college. Among these were the construction of a table showing, as far as the data was available, the number of deliveries and of deaths month by month, and a list of the assistants and students in the sequential order in which they served and practiced in the mater­nity hospital. Professor [Karl] Rokitansky12 has directed the pathological-anatomical division since 1828. From his recollections, and from autopsy reports, and with the help of other physicians and of the assistants and students who participated in the examination of corpses, it would be possible to determine whether the number of diseased patients corresponded to

In consequence of my conviction I must af­firm that only God knows the number of pa­tients who went prematurely to their graves be­cause of me. I have examined corpses to an extent equaled by few other obstetricians. If I say this also of another physician, my intention is only to bring to consciousness a truth that, to humanity’s great misfortune, has remained un­known through so many centuries. No matter how painful and oppressive such a recognition may be, the remedy does not lie in suppression. If the misfortune is not to persist forever, then this truth must be made known to everyone concerned.

After it was realized that the additional deaths in the first clinic were explained by ca­daverous and ichorous particles on the exam­iners’ contaminated hands, various unex­plained phenomena could be accounted for quite naturally. In the morning hours the pro­fessor and the students made general rounds; in the afternoons the assistant and the students made rounds. As part of their instruction, the students examined all patients who were preg­nant or in labor. The assistant was also obliged, before the morning visit of the professor, to examine those in labor and to report on them to the professor. Between these visits the assistant and the students would assume responsibility for necessary examinations. When, therefore, dilation extended over a long period and the patient spent one or more days in the labor room, she was certain to be examined repeat­edly by persons whose hands were contami­nated with cadaverous and ichorous particles. In this way childbed fever was induced, and as I have mentioned, these individuals died almost without exception. Once the chlorine washings were adopted and the patients were examined only by persons with clean hands, patients with extended periods of dilation stopped dying, and extended labor was no more dangerous than in the second clinic.

In order to make my next point intelligible, I must partially explain how I conceive of child-bed fever. For now it is sufficient to observe that foul animal-organic particles are resorbed, and that in consequence of this resorption, disin­tegration of the blood [Blutentmischungl sets in. We have already noted that those with extended periods of dilation contracted rapidly develop­ing childbed fever either during birth or di­rectly thereafter. In other words, the resorption of foul animal-organic particles and the result­ing disintegration of the mother’s blood oc­curred at a time when the fetal blood was in organic exchange through the placenta with the blood of the mother. In this way, blood disin­tegration, from which the mother was suffering, was transmitted to the child. In consequence the newborn, whether female or male, died from a disease identical to that of the mother and in numbers equal to the mothers. Childbed fever originates in the mother because foul ani­mal-organic matter is resorbed and leads to blood disintegration. In the infant the situation is somewhat different. The fetus, as yet unborn and in the birth canal, does not resorb foul animal-organic matter when it is touched by the examiner’s contaminated fingers, but only when its blood is organically mixed with the mother’s blood that has already become con­taminated. This explains why an infant never dies of childbed fever while the mother remains healthy; childbed fever does not arise in the newborn through direct resorption. Both be­come ill while the child and mother are in organic interchange through the placenta and when the blood of the mother has disintegrated through the resorption of foul animal-organic matter. The mother can become ill while the child remains healthy if the organic interchange between them is ended by the birth process before disintegration of the mother’s blood has begun.

As I have said, cadaverous particles adhering to the hands were destroyed by chlorine wash­ings. In this way, the incidence of disease among maternity patients was brought within the limits set in the second clinic. Chlorine washings had the same effect on the incidence of disease among the newborn. Healthy mothers could no longer impart childbed fever to their infants.

In 1846, without chlorine washing, of 3533 infants in the first clinic, 235 died (6 percent). In the second clinic, of 3398 infants 86 died (2.5 percent). In 1847, during the last seven months of which we washed with chlorine, of 3322 in­fants 167 died (5.02 percent). In the second clinic, of 3139 infants 90 died (2.8 percent). In 1848, when chlorine washings were practiced during the entire year, of 3496 infants 147 died in the first clinic (4.2 percent). In the second clinic 100 infants died, out of 3089 (3.2 per­cent). These infant deaths were not from child-bed fever.

If a mother died before her child, or if a mother, for whatever reason, could not nurse her child, the child was taken to the foundling home. In the foundling home, many nursing infants died of childbed fever. After the intro­duction of chlorine washings, nursing infants in the foundling home ceased to die of childbed fever. Dr. [Alois] Bednar, then head physician of the Imperial Foundling Home in Vienna, wrote: “Sepsis of the blood of newborns has become a great rarity. For this we must thank the consequential and most noteworthy discov­ery of Dr. Semmelweis, emeritus assistant of the Viennese first maternity clinic. His work for­tunately explained the cause and the prevention of the formerly murderous ravages of puer­peral fever.”” Where I speak of childbed fever of the newborn, Dr. Bednar correctly speaks of sepsis of the blood; he thus remains consistent with ordinary usage.

Once the cause of the increased mortality in the first clinic was identified as cadaverous par­ticles adhering to the hands of the examiners, it was easy to explain why women who delivered in the street had a strikingly lower mortality rate than those who delivered in the clinic. This was so because once the infant was born and the placenta separated, there was generally no longer opportunity for instruction; thus there were no examinations. A bed was assigned to such patients, and they generally left it in good health. There was no reason for their genitals to be touched by contaminated hands; therefore they did not contract childbed fever. Also, women who delivered prematurely became ill less often because they were not examined either. The first requirement in premature births is to delay birth if possible. Consequently, these persons were not used for open instruction, and decaying organic matter was not conveyed to their genitals.

The sequential appearance of disease was also easy to explain. Because of the large num­ber of births in the first clinic, several indi­viduals were often in the labor room simul­taneously. These persons were examined at least twice a day — during the morning rounds of the professor, and during the afternoon rounds of the assistant. Everyone in labor was examined for instruction sequentially in the order of their beds. When, therefore, the exam­iners’ hands were contaminated with ca­daverous particles, the genitals of several indi­viduals were simultaneously brought into contact with cadaverous particles. This meant that the germ [Keiml for childbed fever was planted through resorption in several indi­viduals at once. The patients were placed back in the maternity ward in the order in which they had delivered. Thus it often happened that those who were together in the labor room de­livered at about the same time and thereafter remained in the same sequential order in the maternity clinic. In the labor room they were examined in rows by persons whose hands were contaminated with cadaverous particles, the germ of the future puerperal fever, and the disease occurred among them sequentially. After chlorine washing was instituted, sequen­tial cases of the disease ceased.

I mentioned that toward the end of 1846, because of the prevalence of childbed fever in the first clinic, yet another commission was in­stituted — I have no idea how many times this had already been done — in order to identify the cause of these deaths. This commission identified the cause as injury to genitals in­flicted during instructional examinations. But because the same examinations were conducted for the instruction of midwives, the commission explained that male students, particularly for­eigners, examined too roughly. Consequently, the number of students was reduced to a mini­mum. Table 3 shows how great the mortality was before this measure was adopted, how it then declined, and how, in the months of April and May, it increased again in spite of the pre­ventive measures. I will now explain these phe­nomena. Before I do, however, one item must be discussed.

As an aspirant for the position of assistant in the first clinic, later as provisional assistant and then, finally, as actual assistant, it was not possi­ble for me to study gynecology at the gynecological division of the Imperial Hospital. However, such study was highly desirable for an obstetrician. As a substitute, as soon as I had decided to devote my life to obstetrics I exam­ined all the female corpses in the morgue of the Imperial General Hospital. From 1844 until I moved to Pest in 1850, I devoted nearly every morning before the professor’s rounds in the obstetrical clinic to these studies. I very much appreciate having enjoyed the friendship of Professor Rokitansky. Through his kindness I secured permission to dissect all female corpses, including those not already set aside for au­topsy, in order to correlate the results of my examinations with autopsies.

For reasons that do not concern us here, the assistant of the first clinic seldom visited the morgue in the months of December 1846 and January, February, and March 1847. The Aus­trian students, whose number was reduced to eighteen, followed his example. The oppor­tunity for them to contaminate their hands with cadaverous particles was thereby greatly re­duced. Restricting examinations to the mini­mum also reduced the opportunity for the geni­tals of patients to be touched by contaminated hands. For these reasons, mortality in the first clinic was reduced during these months.

On 20 March 1847, I reassumed the position of assistant in the first clinic. Early that morning I conducted my gynecological studies in the morgue. I then went to the labor room and began to examine all the patients, as my prede­cessors and I were obliged to do, so that I could report on each patient during the professor’s morning rounds. My hands, contaminated by cadaverous particles, were thereby brought into contact with the genitals of so many women in labor that in April, from 312 deliveries, there were 57 deaths (18.26 percent). In May, from 294 deliveries there were 36 deaths (12.24 per­cent). In the middle of May, without noting the exact day, I instituted chlorine washings. Thus, the great mortality in the first clinic was not caused by injuries in rough examinations — a completely false assumption — but by contami­nated fingers that contacted the genitals of the patients. During April and May, when again so many died, the clinic remained the same as in earlier months, yet the mortality rate increased significantly because I intervened, my fingers contaminated with cadaverous particles.

After chlorine washings were conducted for a longer period with such beneficial results, the number of students was again increased to forty-two. One no longer took account of whether they were Austrian or foreign. The examinations were resumed as was expedient for instruction. Nevertheless, the first clinic lost the dismal distinction of having the greater mortality rate. In December 1846 and in Janu­ary, February, and March of 1847, I functioned as provisional assistant and simultaneously con­ducted gynecological studies in the morgue, yet in these months the mortality rate remained low. The reason is that as provisional assistant I had the right, but not the duty, to examine all patients in labor. After three years in so large a maternity hospital, it was no longer instructive for me to examine all the patients. I examined only exceptional cases — that is, I examined very seldom. When I became the actual assistant, it was my duty to conduct all examinations before the professor’s morning rounds. Thereafter, it was necessary for me to examine nearly all the women in labor for the purpose of instructing the students. This occasioned the great mor­tality rates in April and May of 1847.

Native students are those who completed their education at an Austrian university [Hochschule]. Foreign students are those who were educated elsewhere and who then did fur­ther work at the great University of Vienna. In Vienna one can meet physicians from all the countries of the civilized world. The course in practical obstetrics lasted two months. The in­flux of students into this, the largest maternity hospital in the world, was so great that to accept simultaneously all who sought admission would have excessively disrupted the patients. Appli­cants were assigned numbers, and were ac­cepted sequentially to replace departing stu­dents, regardless of whether they were native or foreign. Each student was free to repeat the course as often as he felt it necessary for his own obstetrical training. However, in order that those who wished to repeat the course would not remain constantly enrolled, precluding oth­ers from taking it at all, it was necessary that one wait three months after completing the course before enrolling again. The commission charged the foreigners with being more dan­gerous than the natives because they were rough in examinations and, consequently, at any one time only two foreigners were allowed to attend the course in practical obstetrics. Everyone, even those who do not share my opinion, will agree that the commission acted groundlessly in imputing guilt to the foreigners. In fact, I alone held that foreigners were more dangerous than natives, but not because they examined more roughly. The reasons that foreigners were more dangerous than natives lies in the following con­siderations.

Foreigners come to Vienna to perfect medical training already begun in their own universities. They visit pathological and forensic autopsies in the general hospital. They take courses in pathological anatomy, in surgery, obstetrics, mi­croscopic surgery of cadavers, they visit the medical and surgical wards of the hospital, etc. In a word, they utilize their time as efficiently and educationally as possible. They have, there­fore, many opportunities for their hands to be­come contaminated with foul animal-organic matter. Thus, it is no wonder that foreigners, busy in the maternity hospital at the same time, are more dangerous for patients. Natives take the course in practical obstetrics after complet­ing two difficult examinations in order to attain the degree of Doctor of Medicine. The law stip­ulates that the minimum preparation time for these examinations is six months. Thus the natives have already toiled excessively before they are admitted into the maternity hospital, and they regard the time there as a rest. While enrolled in practical obstetrics, natives do not concern themselves with other activities that would contaminate their hands. Indeed, while working at the maternity hospital, they concern themselves even less with other aspects of medi­cine because, after completing the course, they can perfect their knowledge of medicine to the highest possible degree. Since the foreigners are generally able to remain in Vienna only a few months, they are compelled to work simul­taneously in more than one aspect of medicine. Even so, one cannot impute guilt to the for­eigners any more than to me or to all the others who undertook examinations with contami­nated hands. None of us knew that we were causing the numerous deaths.